Trackunit launched IrisX MCP today, a new connector that lets ChatGPT, Claude, and Microsoft Copilot tap into live construction equipment data without requiring separate integrations for each assistant. The move, first reported by For Construction Pros, brings conversational AI to heavy machinery fleets, allowing managers to ask anything from “Which generator is running low on fuel?” to “Show me the location of all idle excavators” and get instant answers.
What Trackunit Actually Released
The core of the announcement is IrisX MCP, a server implementation of Anthropic’s Model Context Protocol (MCP). MCP is an open standard that standardizes how AI models retrieve information from external sources. Think of it as a universal adapter: instead of building a custom API client for every AI assistant, a data provider like Trackunit creates one MCP server, and any MCP-compatible AI can connect to it. The IrisX MCP connector exposes a curated set of telematics data from the IrisX platform—things like machine hours, fuel levels, location, fault codes, and maintenance alerts—through this standard interface.
This matters because Trackunit’s IrisX already aggregates data from a wide range of construction equipment OEMs. With the new connector, that aggregated data becomes available inside the AI assistants that office workers, fleet managers, and even field staff are increasingly using. The initial rollout supports ChatGPT, Claude, and Microsoft Copilot. Trackunit’s integration with Copilot is particularly significant for the Windows ecosystem, as it means the very same Copilot that summarizes documents in Word can now also report on excavator health.
What This Means for You
The impact depends on who you are and how you use technology.
For construction fleet managers and operators: If your company already uses Trackunit IrisX to track its machinery, IrisX MCP gives you a new, more natural way to interact with that data. Instead of clicking through dashboards on a laptop or phone, you can open a chat interface—maybe already running on your PC, tablet, or even in Microsoft Teams via Copilot—and type a question. For example: “List all machines that haven’t been serviced in the last 90 days and are currently on site A.” The AI assistant, connected via MCP, will query IrisX and return a formatted list. This can shave minutes off every data lookup, reduce errors from manual dashboard navigation, and make critical information more accessible to non-specialist staff.
For IT administrators and developers in construction firms: This launch simplifies your tech stack. Previously, integrating telematics data into a custom AI agent or enterprise search tool meant building and maintaining bespoke connectors, handling authentication, and keeping up with changing APIs. Now, Trackunit provides the MCP server. You just need to configure your AI client to talk to it. For Microsoft Copilot, this is typically done through Copilot Studio, where you can add an MCP-based knowledge source. You retain control over permissions: the MCP server runs either locally on the user’s machine (keeping data on-premises) or in a cloud deployment, depending on your security policies. This aligns with the zero-trust architectures many enterprises are adopting.
For field workers and mobile users: While MCP setups today are mostly desktop-based, the trend is toward mobile. As AI assistant apps on iOS and Android add MCP support (some already do experimentally), field technicians could query equipment data from their phones using the same chat interface they use for personal tasks. Imagine a technician walking up to a faulty machine, asking Copilot on their phone, “What fault codes are active on this bulldozer, and what are the recommended fixes?” and getting answers drawn from both the machine’s live data and Trackunit’s knowledge base.
For the broader Windows community: This integration is a bellwether for Microsoft’s Copilot strategy. Copilot is evolving from a web-based chatbot into a system-level assistant that can pull data from everywhere—your documents, your calendar, and now, your heavy equipment. For Windows power users and IT pros, it’s a concrete example of how the MCP standard can turn the OS into a unified interface for all your disparate data silos.
How We Got Here
Construction equipment telematics isn’t new. For decades, heavy machinery has been fitted with GPS, engine sensors, and cellular modems. But each manufacturer had its own portal. Trackunit, a Danish company founded in 1998, originally built hardware for asset tracking. Over time, it evolved into a platform that normalizes data from mixed fleets—Caterpillar, Komatsu, Volvo, and others—into one dashboard. IrisX was launched to provide APIs and a data marketplace for that aggregated data.
The missing link was AI. Enterprises wanted to feed that rich, real-time data into the AI assistants their productivity suites were pushing. But every AI assistant had its own extension framework: ChatGPT had plugins, Claude had tool use, Copilot had connectors and plugins. Building for each was costly. Anthropic’s release of the Model Context Protocol in late 2024 changed the game. It offered a common language for data retrieval that any AI model could understand. Microsoft quickly embraced MCP, integrating it into Azure AI and Copilot Studio. Other platform providers followed.
Trackunit’s move was a logical next step. By adopting MCP, it could serve all major AI assistants with one integration. The launch also signals that industrial IoT players are recognizing that their data needs to live inside the tools where decisions are being made—and increasingly, that means inside the chat panes of Copilot, Teams, and Slack.
What You Should Do Now
If you’re a Trackunit customer, here’s how to get started:
- Check your IrisX plan: The MCP connector may require a specific subscription tier. Check with your Trackunit account manager.
- Enable the MCP server: In the IrisX dashboard, look for “Integrations” or “MCP.” You’ll find a server endpoint and credentials to generate.
- Configure your AI assistant:
- For ChatGPT: Use the MCP configuration in the ChatGPT interface to add a new server. Provide the endpoint and API key.
- For Claude: Similar process in the Claude desktop app or API.
- For Microsoft Copilot: The easiest path is through Copilot Studio. Create a new agent, and under “Knowledge,” select “Model Context Protocol.” Enter your IrisX MCP details. Alternatively, developers can use the Azure AI Foundry to build custom copilots that connect via MCP. - Test with simple queries: Start with straightforward prompts like “Show me all active machines” to verify data flow.
- Train your team: Provide examples of powerful queries. The value of natural language data access grows as more teammates learn to ask precise questions.
- Monitor performance and permissions: If you deploy across your organization, audit what data users can access through the AI assistant. MCP connections can be scoped to limit data exposure.
Even if you’re not a Trackunit customer, this launch offers a blueprint. If your organization uses other IoT platforms, ask your vendors whether they support MCP. The protocol is gaining traction, and you may be able to replicate this integration with your own data.
What To Watch Next
The IrisX MCP launch is a stake in the ground for the industrial metaverse: where digital twins and AI meet. Microsoft’s ongoing investment in Copilot and its embrace of MCP suggest that future Windows updates could include native MCP client support, making it even easier to connect to services like IrisX without manual setup. We’ll also be watching for competitors like Samsara, Geotab, or OEM-specific telematics platforms to announce their own MCP connectors. For Windows users, the line between “PC” and “terminal for the physical world” continues to blur, and that’s a trend worth following.